Demonstration of two forms of calcium pumps by thapsigargin inhibition and radioimmunoblotting in platelet membrane vesicles.

In mixed membrane vesicles prepared from human platelets, the presence of two distinct calcium pump enzymes (molecular mass 100 and 97 kDa) was demonstrated by 32P autoradiography, immunoblotting, and thapsigargin inhibition. Both the 100- and 97-kDa membrane proteins showed calcium-dependent phosphoenzyme formation and reacted with a polyclonal anti-sarcoplasmic reticulum calcium pump antiserum, while only the 100-kDa protein reacted with the antiserum specific for the sarco-endoplasmic reticulum-type calcium transport ATPase 2b isoform. Thapsigargin, inhibiting active calcium transport in platelet membrane vesicles, predominantly blocked the phosphoenzyme formation of the 100-kDa isoform and of the tryptic calcium pump fragments of 55 and 35 kDa, while lanthanum specifically increased the phosphoenzyme formation of the 97-kDa enzyme and of the tryptic fragment of 80 kDa. These results indicate the presence of the sarco-endoplasmic reticulum-type calcium transport ATPase 2b isoform and of a yet unidentified, 97-kDa calcium pump protein in human platelet membranes.

In mixed membrane vesicles prepared from human platelets, the presence of two distinct calcium pump enzymes (molecular mass 100 and 97 kDa) was demonstrated by 32P autoradiography, immunoblotting, and thapsigargin inhibition. Both the 100-and 97-kDa membrane proteins showed calcium-dependent phosphoenzyme formation and reacted with a polyclonal anti-sarcoplasmic reticulum calcium pump antiserum, while only the 100-kDa protein reacted with the antiserum specific for the sarco-endoplasmic reticulumtype calcium transport ATPase 2b isoform. Thapsigargin, inhibiting active calcium transport in platelet membrane vesicles, predominantly blocked the phosphoenzyme formation of the 100-kDa isoform and of the tryptic calcium pump fragments of 55 and 35 kDa, while lanthanum specifically increased the phosphoenzyme formation of the 97-kDa enzyme and of the tryptic fragment of 80 kDa. These results indicate the presence of the sarco-endoplasmic reticulum-type calcium transport ATPase 2b isoform and of a yet unidentified, 97-kDa calcium pump protein in human platelet membranes.
Calcium transporting ATPases (Ca2+ pumps) translocate calcium ions through cellular membranes at the expense of ATP hydrolysis to the extracellular space or into intracellular calcium storage sites. Plasma membrane-type Ca2+ pumps are approximately 140-kDa calmodulin-binding proteins encoded by several genes which, by additional alternative splicing, give rise to several structurally related but distinct proteins (1-3).
For the approximately 100-kDa sarco-endoplasmic reticulum-type calcium pumps (SERCA),' three genes have already been identified. The SERCA 1 gene codes for the skeletal muscle sarcoplasmic reticulum calcium-pump; the SERCA l a and Ib are alternatively spliced forms corresponding to the adult and neonatal pump isoforms, respectively. The SERCA *This work has been supported by Grants OTKA-968 and OKKFT-Tt 1.5.1.3. from the Hungarian Academy of Sciences, le Ministere de la Recherche et Technologie, and La Ligue Nationale Contre le Cancer, France. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
11 TO whom correspondence should be addressed U-150 Institut 2 gene also produces alternative spliced species, SERCA 2a and 2b. These are identical, except for a tetrapeptide Cterminal part, present in SERCA 2a, but replaced by a 49amino acid tail in the SERCA 2b isoform. The first one is expressed in heart and slow skeletal sarcoplasmic reticulum, the second is predominantly expressed in smooth muscle and several nonmuscular tissues. A third gene product, SERCA 3, has also been detected by Northern blotting in heart, skeletal muscle, and in a variety of nonmuscular tissues (4-10). The maintenance of low cytosolic calcium concentrations by calcium pumps is essential for the control of platelet activation. Although Ca2+ homeostasis plays an important role in platelet functions, the exact nature of platelet calcium pumps and their regulation is not known. Our previous work, based on the distinct autophosphorylation and transportkinetic characteristics of tryptic fragments, as well as subcellular membrane fractionation and immunoprecipitation experiments indicated the presence of two Ca'+ pump entities in human platelets (11,13,25,(29)(30). However, direct evidence for the presence of two distinct calcium pumps in intact platelet membranes was lacking.
In this work calcium pumps were specifically labeled in intact or trypsinized platelet vesicles by [y3'P]ATP. Upon formation of the aspartylphosphate intermediate of the ATPase, electrophoresis in an acidic polyacrylamide electrophoresis system, previously developed to preserve the phosphoenzyme intermediates (11,12), was carried out. The proteins were blotted onto nitrocellulose, autoradiographed, and immunostained. Since our recent cloning work indicated2 that platelets contain the SERCA 2b isoform, the blots were treated with an antiserum raised against a synthetic peptide, corresponding to the unique C-terminal sequence of this isoform (14). We also examined the immunoreaction with an anti rat-sarcoplasmic reticulum calcium pump antiserum of broad specificity toward calcium pumps of sarco-or endoplasmic reticulum origin (12). In order to differentiate the calcium pump isoforms, the effect of thapsigargin, a recently identified inhibitor of endoplasmic reticulum calcium uptake and tumor promoter (18)(19)(20)(21)23), was studied on the platelet Ca2+ pumps. Inhibition of active calcium uptake into platelet vesicles was measured and compared with the differential inhibitory effect of thapsigargin on the phosphoenzyme formation of the two platelet calcium pumps. In addition, in order to assign the tryptic fragments of the platelet calcium pumps to the respective intact ones, the effect of thapsigargin on the Ca2+ pump fragments in trypsinized platelet membrane Platelet Calcium Pumps vesicles was examined. Moreover, the effect of M e and La3+ on the inhibition by thapsigargin was characterized.
The results obtained indicate the presence of two endoplasmic-reticulum-type Ca2+ pumps in platelets of 100 and 97 kDa apparent molecular mass, respectively, having different autophosphorylation characteristics, thapsigargin sensitivity, and tryptic proteolytic fragmentation pattern.
Platelet membrane vesicles were isolated from fresh platelet-rich plasma, based on the modified version (11) of the method of Kaser-Glanzmann (24). Platelets were washed three times with Tyrode buffer. The washed platelets were resuspended (10" cell/ml) in a solution containing 100 mM KCl, 15 mM NaCl, 2 mM MgC12, 12 mM sodium citrate, 10 mM glucose, 25 mM HEPES-K, pH = 7.5, 0.35% bovine serum albumin, 40 p~ phenylmethylsulfonyl fluoride, 0.1 mM dithiothreitol, and 0.4 mg/ml aprotinin, sonicated three times with an MSE sonicator (maximum amplitude), and centrifuged for 20 min at 35,000 X g at 4 "C. The collected supernatant was sedimented at 16,0000 X g. The pellet was resuspended in 30 mM KCl, 20 mM Tris-HCl, pH = 7.4, and 0.1 mM dithiothreitol, homogenized with a Teflonglass homogenizer, and diluted to a concentration of about 10-15 mg of membrane protein/ml. Calcium Transport Measurements-Ca2+ influx into platelet membrane vesicles was measured for 3 min at 37 "C by the rapid filtration technique described in Ref. 27. The transport medium contained 130 mM KC1,34 mM HEPES-K, pH = 7.2,2 mM MgC12,5 mM K-oxalate, 100 p~ CaC12 (labeled with 45Ca2+), 110 p~ EGTA (free Ca" concentration being 1.6 p~) and 140 or 500 pg of membrane protein/ml, as indicated. Platelet membrane vesicles were preincubated with the required amount of thapsigargin for 3 min at 37 "C before initiating calcium uptake by the addition of 0.5 mM ATP.
Limited proteolysis of platelet membranes was carried out on ice in the presence of 50 pg/ml trypsin (11). The proteolysis medium contained 130 mM KCl, 34 mM HEPES-K, pH = 7.0, 0.5 mM dithiothreitol, 50 p~ CaC12, and 0.5-2 mg of membrane protein/ml. After 10 min of digestion, the reaction was stopped with 5-fold excess of soybean trypsin inhibitor. Membrane phosphorylation experiments were performed immediately after proteolysis.
Phosphorylation of the Calcium Pump($ in Platelet Membrane Vesicles-Phosphorylation was carried out on ice in the medium used for proteolysis, containing 50 p M CaC12, 100 p M LaCln, 1 mM MgCI2, and thapsigargin as indicated. The reaction was started by the addition of 0.05 p~ ATP (including [Y-~~PIATP). After 1 min of incubation, the reaction was stopped by the addition of ice-cold trichloroacetic acid (6% final concentration) containing 1 mM ATP and 10 mM phosphoric acid, and the samples were washed three times with the same solution. The precipitates were dissolved in the electrophoresis sample buffer and either counted for radioactivity or analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis as described in Ref. 28.
Electrophoresis and Immunoblotting-Acidic polyacrylamide gel electrophoresis and Western blotting was performed (12,28) with the following modifications: electrophoresis was done in a Bio-Rad Protean 16CM electrophoresis unit at 4 "C at 120 V constant voltage. Sample buffer contained 150 mM Tris, 8 mM EDTA-Na, 3% (w/v) sodium dodecyl sulfate, 20% (w/v) Saccharose, 0.14 mg/ml bromphenol blue, 1% dithiothreitol, and was adjusted to pH = 7.4 with HCI. After dissolution of the trichloroacetic acid-precipitated protein pellet of sample buffer, samples were realkalinized with 1.7 M Tris, and 2 0 4 samples, corresponding to 100-300 pg of membrane protein, were deposited per well. After electrophoresis and immunoblotting, membranes were autoradiographed using KODAK X-OMAT AR films.
Immunostaining was done by an avidin-biotin-peroxidase staining system as in Ref. 22. For staining with the SERCA 2b antiserum, membranes were quenched in 6 mM Tris, 0.5 M NaCl, 5% (w/v) bovine serum albumin, pH = 7.4, supplemented with 20% heatinactivated human serum, added to increase specificity of immunostaining. The anti-SERCA 2b rabbit antiserum was diluted 120-fold in the same solution.

Electrophoretic Resolution of Two Cu2' Pumps in Platelet
Membrane Vesicles-Ca2+ pumps in human platelet membrane vesicles were labeled by [y3'P]ATP in the presence of Ca2+. After electrophoresis of the membrane proteins and blotting onto nitrocellulose, two closely migrating but distinct phosphorylated bands of molecular mass of about 100 and 97 kDa, respectively, were obtained upon autoradiography (Fig.   1, punel A, (lanes 1-4) head) was more pronounced (Fig. 1, panel A , lune 1 ). Addition of 100 p~ La3+ during phosphorylation resulted in an increase of radiolabeling of the lower band (97 kDa, lune 2, open arrowhead), while it did not affect the labeling of the 100-kDa protein.

Different Reactivity of the Two Ca2+ Pumps with Anti-Ca"
Pump Antibodies-Both bands were recognized upon immunostaining (Fig. 1, panel B ) by a polyclonal antiserum (lune 3 ) , raised against rat sarcoplasmic reticulum Ca2+ pump, whereas only the upper (100 kDa) band was stained by an antiserum, specific for the SERCA 2b isoform (lane 4 ) . Thus, two distinct Ca2+ pump proteins can be discerned in platelet membrane vesicles, having different molecular masses, lanthanum sensitivities of phosphoenzyme formation, and immunoreactivities toward anti-Ca2+ pump antibodies.
To further characterize these two calcium transport systems, the effect of thapsigargin on the phosphoenzyme formation of the Ca2+ pumps and on the ATP-dependent Ca2+ uptake into platelet membrane vesicles was studied.
Inhibition by Thapsigargin of Active Ca2+ Uptake into Platelet Membrane Vesicles-As shown on Fig. 2, thapsigargin inhibited active Ca2+ uptake into platelet membrane vesicles. Inhibition by thapsigargin appeared to be dependent on the inhibitor to membrane protein ratio, half-maximal inhibition obtained at 120-140 nM at 0.5 mg/ml, and at 35-40 nM thapsigargin a t 0.14 mg/ml membrane protein concentration, respectively. ICso was about 250 pmol of thapsigargin/mg of membrane protein.

Inhibition by Thapsigargin of Phosphoenzyme Intermediate Formation of Ca2+ Pumps in Platelet Membrane Vesicles-As
shown on Fig. 3, thapsigargin also caused a dose-dependent inhibition of total phosphoenzyme formation, with a halfmaximal concentration of about 100 nM at 0.4 mg/ml membrane protein concentration, corresponding again to an ICs0 of about 250 pmol of thapsigargin/mg of membrane protein. This inhibition was more pronounced in the presence of Mg?+ plus Ca" than with Ca2+ alone and abolished by La'+ in the presence, as well as in the absence of M e . Since M$+ decreases steady state phosphoenzyme level by facilitating the hydrolysis of the aspartylphosphate group, it can be concluded that thapsigargin interferes with the phosphorylation step of the transport cycle, rather than inhibiting the dephosphorylation step, as in the case of lanthanum, which blocks the enzyme in the phosphorylated state. Fig.   4 shows the effect of thapsigargin on the appearance of the phosphorylated bands after the membrane proteins were resolved by acidic polyacrylamide gel electrophoresis. Thapsigargin inhibited the phosphoenzyme formation predominantly in the case of the higher molecular mass (100 kDa) enzyme (SERCA 2b). While in the absence of thapsigargin under the reaction conditions used (i.e. in the presence of Ca2+ and M e , without La", a t 0.4 mg of membrane protein/ ml) the upper band appears to be more phosphorylated (Fig.  4. lane 1 ), the addition of thapsigargin results in an almost complete disappearence of this band in the submicromolar thapsigargin concentration range (Fig. 4. lanes 2-4); at 2 pM thapsigarin concentration practically only the lower 97-kDa pump species remained phosphorylated, although to a lesser extent, than in the absence of the inhibitor. At this thapsigargin concentration, the Ca2+-uptake is almost completely inhibited. This slight discrepancy between transport and phosphorylation experiments is probably due to the different assay conditions used in the two types of experiments (see "Experimental Procedures").

Assignment of Tryptic Fragments to the Respective Nonproteolysed Ca2' Pumps by Differential La" and Thapsigargin
Sensitivity-Upon mild trypsin proteolysis (Fig. 4, panel B ) , platelet membrane vesicles give rise to 80-, 55-, and 35-kDa autophosphorylable Ca2+ pump fragments (11). Out of these fragments, the autophosphorylation of the 80-kDa fragment is enhanced by lanthanum ( l l ) , similarly to that observed in the case of the 97-kDa nonproteolyzed pump, in this work (Fig. 1, lune 2). This suggests that the 80-kDa fragment may be the cleavage product of the 97-kDa Ca'+ pump species, the 55and 35-kDa fragments coming from the 100 kDa form. In order to test this hypothesis, platelet membrane vesicles were subjected to limited trypsin proteolysis, and thapsigargin inhibition of the phosphoenzyme formation of the Ca'+ pump fragments was studied. As seen on Fig. 4, lane 7, at 200 nM thapsigargin concentration the phosphoenzyme level of the 55-kDa Ca2+ pump fragment was decreased, whereas that of the 80-kDa fragment remained unchanged, as compared to the control (lane 6). The autophosphorylation of the 35-kDa fragment was also inhibited by 200 nM thapsigargin (not shown). The same pattern of inhibition of phosphoenzyme formation was obtained for the non-proteolyzed Ca2+ pumps, the 100 kDa form being inhibited, and the 97 kDa form being relatively insensitive to this thapsigargin concentration (Fig.   4. lane 4). At 2 PM thapsigargin both fragments (80 and 55 kDa) appeared to be inhibited (Fig. 4, lane 8 ) , although to a lesser extent than the respective nonproteolyzed ones, probably due to differences in the affinity of thapsigargin to the more or less truncated pump fragments. Addition of lanthanum abolished thapsigargin inhibition of calcium pump autophosphorylation in the case of proteolyzed, as well as nonproteolyzed platelet membranes (Fig 4, lanes 5 and 9).

DISCUSSION
In the present work direct evidence is presented for the existence of two separate calcium transporting ATPases in human platelets. Two closely migrating endoplasmic reticulum-type pump species were resolved on polyacrylamide gels and identified by their capacity of phosphoenzyme intermediate formation and reactivity with anti-Ca2+ pump antibodies. This approach has already been widely used in the study of calcium pumps in different cell types (11,12,22), including platelets, with success for the resolution of calcium pump fragments after trypsinolysis (11,12).
The higher molecular weight band is identified as the SERCA 2b isoform, recently cloned from platelet libraries in our laboratory.' As shown in this work, it has high thapsigargin sensitivity and reacts with an antibody specific for the SERCA 2b form.
The other form, with 97-kDa apparent molecular mass in our gel system, was different from the SERCA 2b form with respect to its increase of steady state phosphoenzyme level in the presence of lanthanum, lower sensitivity to thapsigargin inhibition, and lack of reactivity toward the anti-SERCA 2b antiserum, although it was well stained by a polyclonal antiserum raised against the rat sarcoplasmic reticulum Ca'+ Pump.
In platelet membrane vesicles upon mild trypsin proteolysis 80-, 5 5 , and 35-kDa autophosphorylable Ca2+ pump fragments are produced (11). The 80-kDa tryptic fragment is similar to the 97-kDa nonproteolyzed form, in terms of enhancement of the phosphoenzyme formation by lanthanum and low sensitivity to thapsigargin inhibition. On the other hand, the 55-and 35-kDa pump fragments resemble the 100-kDa SERCA 2b isoform; their phosphoenzyme formation is unaffected by lanthanum and inhibited by thapsigargin. Based on these data, in platelets the 80-kDa tryptic Ca2+ pump fragment can be assigned to the 97-kDa intact Ca2+ pump, whereas the 55-and 35-kDa fragments are most probably cleavage products of the 100-kDa SERCA 2b isoform. Al-though the enhancement of the phosphoenzyme level by lanthanum and the formation of a 80-kDa tryptic fragment are usually found in the case of plasma membrane-type (140 kDa) Ca'+ pumps, the presence of such a pump in platelets can be ruled out, since no 140-kDa phosphorylated Ca'+ pump was found in platelet membranes and plasma membrane-type calcium transport ATPase-specific mono-or polyclonal antibodies did not stain platelet membrane vesicle proteins on Western blots (not shown). The 80-, as well as the 55-, and 35-kDa phosphorylated pump fragments are recognized by the anti-sarcoplasmic reticulum Ca'+ pump antiserum (12). On the other hand, as expected, the anti-SERCA 2b antiserum, raised against a C-terminal oligopeptide, failed to stain any of these pump fragments (not shown).
In addition, although sarcoplasmic and plasma membrane Ca'+ pumps are reportedly unaffected by thapsigargin (21), our data confirm that various endoplasmic reticulum Ca'+ pumps may also have different sensitivities toward thapsigargin inhibition.